tiny fix

examples/extras/fsdp_qlora/single_node.sh

@@ -1,10 +1,6 @@
 #!/bin/bash
 # DO NOT use GPTQ/AWQ model in FSDP+QLoRA
 
-pip install "transformers>=4.39.1"
-pip install "accelerate>=0.28.0"
-pip install "bitsandbytes>=0.43.0"
-
 CUDA_VISIBLE_DEVICES=0,1 accelerate launch \
     --config_file examples/accelerate/fsdp_config.yaml \
     src/train.py examples/extras/fsdp_qlora/llama3_lora_sft.yaml